The current bridge maintenance practice generally involves manual visual inspection, which is highly subjective and unreliable. A technique that can automatically detect defects, for example, surface cracks, is essential so that early warnings can be triggered to prevent disaster due to structural failure. In this study, to permit automatic identification of concrete cracks, an ad‐hoc faster region‐based convolutional neural network (faster R‐CNN) was applied to contaminated real‐world images taken from concrete bridges with complex backgrounds, including handwriting. A dataset of 5,009 cropped images was generated and labeled for two different objects, cracks and handwriting. The proposed network was then trained and tested using the generated image dataset. Four full‐scale images that contained complex disturbance information were used to assess the performance of the trained network. The results of this study demonstrate that faster R‐CNN can automatically locate crack from raw images, even with the presence of handwriting scripts. For comparative study, the proposed network is also compared with You Only Look Once v2 detection technique.

An algorithm based on correlation analysis was adopted to estimate the probability of the presence of damage in aluminum plates using Lamb wave signals from an active sensor network. Both finite element analysis and experimental evaluations were presented. The Shannon entropy optimization criterion was applied to calibrate the optimal mother wavelet and the most appropriate continuous wavelet transform scale for signal processing. The correlation coefficients for individual sensing paths between the present state (with damage) and the reference state (without damage) were calculated, and the probability of the presence of damage in the monitoring area enclosed by the active sensor network was estimated to identify the damage. A concept of virtual sensing paths (VSPs) was proposed to enhance the performance of the algorithm by increasing the number of sensing paths in data fusion. The results identified using both simulation and experimental Lamb wave signals from different groups of sensing paths at different central frequencies agreed well with the actual situations, demonstrating the potential of the correlation-based algorithm with the application of VSPs for identification of damage in structures.